Short-term PV power forecasting aims to obtain complex information features from historical data to predict data for a short interval in the future. This task is often used to control system operation and fault detection. However, PV power data exhibit high variability, and large-scale power fluctuations cannot be adapted to by the combined model when predicting, thus falling into a local optimal solution. This article proposes a dynamic combination of the TCN-BiGRU and TCN-BiLSTM short-term solar power forecasting models based on CEEMDAN. The volatility of the primary data can be alleviated by this approach while preventing the model from falling into a local optimal solution. Experiments show that this method has reliable precision and flexibility, which suitable for different short-term PV power prediction scenarios.

Accurate short-term photovoltaic (PV) power prediction can be crucial for fault detection of the control system and reducing the fault of the PV output control system. However, PV power is highly volatile, and significant power fluctuations cannot be adapted to by the combined model when predicting, thus affecting the stable operation of the PV output control system. In response to this issue, a dynamic combination short-term PV power prediction model of temporal convolutional network (TCN)-bidirectional gated recurrent unit network (BiGRU) and TCN-bidirectional long-short term memory network (BiLSTM) based on complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) is proposed. CEEMDAN is employed to decompose the original PV power data to reduce the volatility of the original data. Constructing two combined models, TCN-BiGRU and TCN-BiLSTM, and training them separately. Introducing ElasticNet, which utilizes both L1 and L2 regularization terms. This approach prese...